[ghorrocks]

If your flow is near turbulence transition it is going to be very tricky. Normal turbulence models do not work well in this region (but some work better than others - SST is far better than normal k-e for instance).

You are probably going to need a transient LES style simulation, and a laminar flow model may well be as good as anything. For example my PhD thesis looked at exactly this sort of thing in a piston/cylinder arrangement - see http://hdl.handle.net/10453/20133 in chapter 6.

[fkhan7]

I am trying to simulate thermocline formation in a tank and so far i have tried K-e and SST model with no luck of matching my experimental results. I will try LES and post my results.

Thank you for your input and sharing your thesis,

Fahad

[FluidCFD]

Hi,

I have successfully done my simulation for a 6.5 m³ tank with a stratification device with diameter 125 mm and a volume flow of 1 m³/h. I used laminar flow and the results match the experimental data quite well (see attachment).

The green line is the starting temperature distribution, the yellow line the inflow temperature. The purple curve shows the result of the experiment (after 3 h 7 min). Finally, the red line is my simulation result.

I'm still not sure how to proceed with higher volume flows. If you have new insights concerning this problem, please let me know.

Best regards.
G.

[fkhan7]

Hello G,

what is you Re and Fr number and what time of inlet shape are you using?

From what i have seen now, even at Re= 1000 the flow seem to be turbulent.

Regards,

Fahad

[FluidCFD]

Hi,

the type of inlet shape can be seen in the attachments of post #7 and #17 in this thread. What dimensions do you use in Re and Fr? Diameter of stratifier or tank or some other dimensions?

Best regards.
G.

[farah]

Hi all
i want to generate 3 d triplex pipe by workbench but i have problem with faces generation
any help please